hashirama/uvg266
1
0
Fork 0
mirror of https://github.com/ultravideo/uvg266.git synced 2024-11-27 11:24:05 +00:00
Code Issues Actions Projects Releases Wiki Activity

[avx2] static all transform tables

Browse source
This commit is contained in:
Joose Sainio 2023-07-26 15:24:31 +03:00
parent 1f9955bdda
commit d62a3f888e
2 changed files with 254 additions and 261 deletions
Show stats Download patch file Download diff file Expand all files Collapse all files

96
src/strategies/avx2/dct-avx2.c
View file

@ -82,7 +82,7 @@ static INLINE __m256i truncate_avx2(__m256i v, __m256i debias, int32_t shift)
// TODO: find avx2 solution for transpose
// TODO: attempt to make a generic transpose for avx2. Needs some extra logic for different widths and heights.
// TODO: make a few solutions for exact sizes and see if some pattern emerges...
void transpose_matrix(const int16_t* src, int16_t* dst, const int width, const int height) {
static void transpose_matrix(const int16_t* src, int16_t* dst, const int width, const int height) {
const int sample_num = width * height;
const int vectors = sample_num / 16;
@ -150,7 +150,7 @@ void transpose_matrix(const int16_t* src, int16_t* dst, const int width, const i
}
}
void transpose_generic(const int16_t* src, int16_t* dst, const int width, const int height)
static void transpose_generic(const int16_t* src, int16_t* dst, const int width, const int height)
{
for (int y = 0; y < height; ++y) {
for (int x = 0; x < width; ++x) {
@ -644,7 +644,7 @@ static transpose_func* transpose_func_table[6][6] = {
// Dispatcher function for avx2 transposes. This calls the proper subfunction
void transpose_avx2(const __m256i* src, __m256i* dst, const int width, const int height)
static void transpose_avx2(const __m256i* src, __m256i* dst, const int width, const int height)
{
// No need to transpose something of width or height 1
const int w_log2_minus1 = uvg_g_convert_to_log2[width] - 1;
@ -2043,7 +2043,7 @@ static void fast_forward_tr_2xN_avx2_hor(const int16_t* src, __m256i* dst, const
}
}
void fast_forward_tr_2x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_2x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 2;
const int height = 8;
@ -2162,7 +2162,7 @@ static void fast_inverse_tr_2x8_avx2_hor(const __m256i* src, int16_t* dst, const
_mm256_store_si256((__m256i*)dst, v_result);
}
void fast_inverse_tr_2x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_2x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 2;
const int height = 8;
@ -2187,7 +2187,7 @@ void fast_inverse_tr_2x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, t
}
void fast_forward_tr_2x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_2x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 2;
const int height = 16;
@ -2350,7 +2350,7 @@ static void fast_inverse_tr_2x16_avx2_hor(const __m256i* src, int16_t* dst, cons
_mm256_store_si256((__m256i*) & dst[16], v_result_1);
}
void fast_inverse_tr_2x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_2x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 2;
const int height = 16;
@ -2375,7 +2375,7 @@ void fast_inverse_tr_2x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor,
}
void fast_forward_tr_2x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_2x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 2;
const int height = 32;
@ -2551,7 +2551,7 @@ static void fast_inverse_tr_2x32_avx2_hor(const __m256i* src, int16_t* dst, cons
}
}
void fast_inverse_tr_2x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_2x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 2;
const int height = 32;
@ -2574,7 +2574,7 @@ void fast_inverse_tr_2x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor,
}
void fast_forward_tr_4xN_avx2_hor(const int16_t* src, __m256i* dst, const int16_t* coeff, int32_t shift, int line, int skip_line, int skip_line2)
static void fast_forward_tr_4xN_avx2_hor(const int16_t* src, __m256i* dst, const int16_t* coeff, int32_t shift, int line, int skip_line, int skip_line2)
{
const int32_t add = (shift > 0) ? (1 << (shift - 1)) : 0; // ISP_TODO: optimize (shift > 0) check out if shift is always gt 0
const __m256i debias = _mm256_set1_epi32(add);
@ -2620,7 +2620,7 @@ void fast_forward_tr_4xN_avx2_hor(const int16_t* src, __m256i* dst, const int16_
}
}
void fast_forward_tr_4x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_4x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 4;
const int height = 4;
@ -2733,7 +2733,7 @@ static void fast_inverse_tr_4x4_avx2_ver(const __m256i* src, int16_t* dst, const
_mm256_store_si256((__m256i*)dst, v_result);
}
void fast_inverse_tr_4x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_4x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 4;
const int height = 4;
@ -2764,7 +2764,7 @@ void fast_inverse_tr_4x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, t
}
void fast_forward_tr_4x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_4x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 4;
const int height = 8;
@ -2941,7 +2941,7 @@ static void fast_inverse_tr_4x8_avx2_hor(const __m256i* src, int16_t* dst, const
_mm256_store_si256((__m256i*) & dst[16], v_result_1);
}
void fast_inverse_tr_4x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_4x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 4;
const int height = 8;
@ -2972,7 +2972,7 @@ void fast_inverse_tr_4x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, t
}
void fast_forward_tr_4x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_4x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 4;
const int height = 16;
@ -3169,7 +3169,7 @@ static void fast_inverse_tr_4x16_avx2_hor(const __m256i* src, int16_t* dst, cons
}
}
void fast_inverse_tr_4x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_4x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 4;
const int height = 16;
@ -3200,7 +3200,7 @@ void fast_inverse_tr_4x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor,
}
void fast_forward_tr_4x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_4x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 4;
const int height = 32;
@ -3404,7 +3404,7 @@ static void fast_inverse_tr_4x32_avx2_hor(const __m256i* src, int16_t* dst, cons
}
}
void fast_inverse_tr_4x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_4x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 4;
const int height = 32;
@ -3495,7 +3495,7 @@ static void fast_forward_tr_8xN_avx2_hor(const int16_t* src, __m256i* dst, const
}
}
void fast_forward_tr_8x2_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_8x2_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 8;
const int height = 2;
@ -3613,7 +3613,7 @@ static void fast_inverse_tr_8x2_avx2_hor(const __m256i* src, int16_t* dst, const
_mm256_store_si256((__m256i*)dst, v_result);
}
void fast_inverse_tr_8x2_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_8x2_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 8;
const int height = 2;
@ -3637,7 +3637,7 @@ void fast_inverse_tr_8x2_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, t
fast_inverse_tr_8x2_avx2_hor(&v_ver_pass_out, dst, hor_coeff, shift_2nd, height, 0, skip_width);
}
void fast_forward_tr_8x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_8x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 8;
const int height = 4;
@ -3804,7 +3804,7 @@ static void fast_inverse_tr_8x4_avx2_hor(const __m256i* src, int16_t* dst, const
_mm256_store_si256((__m256i*) & dst[16], v_result[1]);
}
void fast_inverse_tr_8x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_8x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 8;
const int height = 4;
@ -3835,7 +3835,7 @@ void fast_inverse_tr_8x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, t
}
void fast_forward_tr_8x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_8x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 8;
const int height = 8;
@ -4052,7 +4052,7 @@ static void fast_inverse_tr_8x8_avx2_ver(const __m256i* src, int16_t* dst, const
}
}
void fast_inverse_tr_8x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_8x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 8;
const int height = 8;
@ -4083,7 +4083,7 @@ void fast_inverse_tr_8x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, t
}
void fast_forward_tr_8x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_8x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 8;
const int height = 16;
@ -4331,7 +4331,7 @@ static void fast_inverse_tr_8x16_avx2_hor(const __m256i* src, int16_t* dst, cons
}
}
void fast_inverse_tr_8x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_8x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 8;
const int height = 16;
@ -4362,7 +4362,7 @@ void fast_inverse_tr_8x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor,
}
void fast_forward_tr_8x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_8x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 8;
const int height = 32;
@ -4576,7 +4576,7 @@ static void fast_inverse_tr_8x32_avx2_hor(const __m256i* src, int16_t* dst, cons
// TODO: mts cutoff
}
void fast_inverse_tr_8x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_8x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 8;
const int height = 32;
@ -4689,7 +4689,7 @@ static void fast_forward_DCT2_B16_avx2_hor(const int16_t* src, __m256i* dst, con
}
}
void fast_forward_tr_16x2_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_16x2_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 16;
const int height = 2;
@ -4806,7 +4806,7 @@ static void fast_inverse_tr_16x2_avx2_hor(const __m256i* src, int16_t* dst, cons
_mm256_store_si256((__m256i*) & dst[16], v_result_1);
}
void fast_inverse_tr_16x2_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_16x2_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 16;
const int height = 2;
@ -4831,7 +4831,7 @@ void fast_inverse_tr_16x2_avx2(const int16_t* src, int16_t* dst, tr_type_t hor,
}
void fast_forward_tr_16x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_16x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 16;
const int height = 4;
@ -5039,7 +5039,7 @@ static void fast_inverse_tr_16x4_avx2_hor(const __m256i* src, int16_t* dst, cons
}
}
void fast_inverse_tr_16x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_16x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 16;
const int height = 4;
@ -5070,7 +5070,7 @@ void fast_inverse_tr_16x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor,
}
void fast_forward_tr_16x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_16x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 16;
const int height = 8;
@ -5293,7 +5293,7 @@ static void fast_inverse_tr_16x8_avx2_hor(const __m256i* src, int16_t* dst, cons
}
}
void fast_inverse_tr_16x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_16x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 16;
const int height = 8;
@ -5324,7 +5324,7 @@ void fast_inverse_tr_16x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor,
}
void fast_forward_tr_16x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_16x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 16;
const int height = 16;
@ -5629,7 +5629,7 @@ static void fast_inverse_tr_16x16_avx2_ver(const __m256i* src, int16_t* dst, con
//transpose_avx2(v_result, (__m256i*)dst, 16, 16);
}
void fast_inverse_tr_16x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_16x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 16;
const int height = 16;
@ -5660,7 +5660,7 @@ void fast_inverse_tr_16x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor,
}
void fast_forward_tr_16x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_16x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 16;
const int height = 32;
@ -5938,7 +5938,7 @@ static void fast_inverse_tr_16x32_avx2_hor(const __m256i* src, int16_t* dst, con
}
}
void fast_inverse_tr_16x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_16x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 16;
const int height = 32;
@ -6285,7 +6285,7 @@ static void fast_forward_DCT2_32x8_avx2_ver(const __m256i* src, int16_t* dst, in
}
void fast_forward_tr_32x2_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_32x2_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 32;
const int height = 2;
@ -6450,7 +6450,7 @@ static void fast_inverse_tr_32x2_avx2_hor(const __m256i* src, int16_t* dst, cons
// TODO: cutoff for DCT8 and DST7
}
void fast_inverse_tr_32x2_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_32x2_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 32;
const int height = 2;
@ -6472,7 +6472,7 @@ void fast_inverse_tr_32x2_avx2(const int16_t* src, int16_t* dst, tr_type_t hor,
}
void fast_forward_tr_32x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_32x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 32;
const int height = 4;
@ -6879,7 +6879,7 @@ static void fast_inverse_tr_32x4_avx2_mts_hor(const __m256i* src, int16_t* dst,
// TODO: cutoff for dct8 and dst7
}
void fast_inverse_tr_32x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_32x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 32;
const int height = 4;
@ -6920,7 +6920,7 @@ void fast_inverse_tr_32x4_avx2(const int16_t* src, int16_t* dst, tr_type_t hor,
}
void fast_forward_tr_32x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_32x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 32;
const int height = 8;
@ -7206,7 +7206,7 @@ static void fast_inverse_tr_32x8_avx2_hor(const __m256i* src, int16_t* dst, cons
// TODO: cutoff for dct8 and dst7
}
void fast_inverse_tr_32x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_32x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 32;
const int height = 8;
@ -7242,7 +7242,7 @@ void fast_inverse_tr_32x8_avx2(const int16_t* src, int16_t* dst, tr_type_t hor,
}
void fast_forward_tr_32x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_32x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 32;
const int height = 16;
@ -7615,7 +7615,7 @@ static void fast_inverse_tr_32x16_avx2_hor(const __m256i* src, int16_t* dst, con
// TODO: MTS cutoff
}
void fast_inverse_tr_32x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_32x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 32;
const int height = 16;
@ -7646,7 +7646,7 @@ void fast_inverse_tr_32x16_avx2(const int16_t* src, int16_t* dst, tr_type_t hor,
}
void fast_forward_tr_32x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_forward_tr_32x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 32;
const int height = 32;
@ -8012,7 +8012,7 @@ static void fast_inverse_tr_32x32_avx2_hor(const __m256i* src, int16_t* dst, con
}
}
void fast_inverse_tr_32x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
static void fast_inverse_tr_32x32_avx2(const int16_t* src, int16_t* dst, tr_type_t hor, tr_type_t ver)
{
const int width = 32;
const int height = 32;

419
src/strategies/avx2/dct_avx2_tables.h
View file

File diff suppressed because it is too large Load diff